Jensen Huang resigns: Colette Kress is Nvidia's interim CEO

Share
Jensen Huang resigns: Colette Kress is Nvidia's interim CEO

Nvidia's board announced Thursday that Jensen Huang has resigned as president, chief executive and a member of the board, effective immediately, the company said in a statement. Colette Kress, chief financial officer since 2013, becomes interim CEO and keeps the CFO job. Harvey C. Jones, the former Synopsys chief executive, was appointed the first chair of the board in Nvidia's 33-year history.

The statement carries two quotes: one from Jones, one from Kress. Neither is from Huang. There is no transition period, no advisory role, no chairman emeritus, and the thanks to the founder — "Jensen's contributions to NVIDIA are profound" — is written in the third person about a man who left the building the same morning. Boards do not phrase it that way when they expect the founder to stay involved.

The market read it as a removal. Shares fell about 12% in premarket trading and were roughly 6% lower shortly after the open, per Stock Analysis's quote. CoreWeave, the neocloud that leases Nvidia capacity and counts Nvidia as an investor, dropped 6% after hours on Wednesday, Reuters reported.

The chair is the real structural change

Nvidia had 11 directors and considered all but Huang independent, Reuters reported this week. It had no chair, because the CEO ran the board. He also had no president, no chief operating officer and around 60 direct reports — the flat structure we described when the board first opened its investigation.

Appointing Jones changes the shape of that. Jones ran Synopsys, which sells the design software every chip company on earth, Nvidia included, builds on. He is a semiconductor-toolchain operator, not an AI person, and he is now the board's chair rather than a figurehead the CEO appoints. That matters for one reason: it moves the succession decision out of the executive suite and into a room where the interim CEO has to report.

It also makes an outside hire more plausible. A board that wanted an internal caretaker would have given Kress the chair too, or named her permanent CEO on day one and ended the story. Instead the statement says the board "is initiating a search for a permanent Chief Executive Officer and will consider internal and external candidates" — language that keeps the job open for someone who has never worked at Nvidia.

Five days to an earnings call the CFO now answers alone

Nvidia reports third-quarter fiscal 2027 results on Tuesday, November 17, as previously scheduled. The company guided to $108 billion for the quarter, with an explicit assumption of no data-centre compute revenue from China in the number. Kress has spent thirteen years explaining that arithmetic to analysts; on Tuesday she explains it as the person in charge of the company that produced it.

The harder problem is what sits outside the income statement. Nvidia holds roughly $99 billion of stakes in AI companies, from frontier labs to neoclouds, The Verge reports, and it has committed to backstop part of a financing structure with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR aimed at mobilising more than $500 billion. Those agreements were negotiated by Huang personally, in most cases with the counterparty's CEO across the table.

An interim chief executive who is also the chief financial officer has one instinct, and it is the correct one: do not sign anything new that has to be marked to market before the auditors and the permanent CEO arrive. Watch for a quiet freeze on new strategic cheques. That is a bigger deal for the AI buildout than the chip roadmap, because Nvidia's investment book has become the down payment on other companies' capacity, and a pause at the centre of it lands on the customers who were counting on the next commitment.

The China argument lost its author

Huang's second job was salesman for a policy. He argued, publicly and repeatedly, that export controls cost American chipmakers market share and pushed China to build its own accelerators, and he took that case to President Trump in person. Whatever the merits, the argument had a face and a phone number. The interim CEO inherits the position without the relationship.

Two dates will show whether the company keeps making that case. Huang was due to appear before three Republican committee chairs for a closed-door briefing on November 18, the day after earnings. Someone from Nvidia still has to sit in that chair, and the answer — counsel, the new chair, or Kress herself — will say how much legal exposure the company thinks it has.

The Justice Department thread is the one that outlives the personnel story. WIRED reports that Commerce's investigation into Megaspeed, the Singapore cloud provider where Alice Huang held a senior role, was halted and transferred to Justice after Nvidia's general counsel inquired about it. A criminal inquiry does not care who the CEO is. It cares what the company knew about a customer buying billions of dollars of its hardware through related entities.

Why the board moved, and why the obvious hire is awkward

WIRED's source says the relationship with a direct report, plus the Megaspeed dealings, probably weighed more in the board's decision than the espionage allegations themselves. That fits the sequence: the board publicly backed Huang on Sunday, then reversed within 72 hours of the second disclosure. A relationship with a subordinate is a documented policy question the board can adjudicate on its own paper. A foreign intelligence assessment is not.

The search has no obvious internal answer, which is why the names circulating are external. The Verge reports that Apple's hardware chief Johny Srouji has been floated at the top of the list. Srouji would be a strange fit in one respect and an obvious one in another: he has shipped silicon at scale for two decades, and he has never had to allocate scarce supply among five frontier labs that all consider themselves the priority customer.

The contrarian case

The bull argument is that demand is booked, CUDA is code rather than a person, and customers will buy the same number of GPUs with or without the founder. It is a serious argument. Nvidia's backlog and its guidance were set before this week, and no purchase order was cancelled on Thursday.

The honest counter is that the part of Huang's job that has no deputy was coordination: deciding who gets allocation in a shortage, pricing the systems business against the cloud renters, keeping TSMC and the memory makers on the same cadence, and making the personal call that unlocked a commitment. None of that appears in an org chart, and all of it now runs through a finance executive on an interim basis and a chair who last ran a software company.

What to watch

Tuesday's call, first. Whether Kress repeats the investment commitments in the same language, whether she takes questions on the portfolio, and whether guidance for the January quarter keeps the no-China assumption. Then November 18, and who Nvidia sends to Congress. Then the search: a permanent CEO named before the fiscal year ends in January would signal the board wants distance from the scandal; a search that drags into spring signals it could not find anyone willing to take a 60-direct-report job with a subpoena attached.

The stock has lost about $600 billion in market value since the board announced its investigation, Reuters reported. The question for the next month is not whether Nvidia can sell GPUs without Huang. It is whether the web of investments, backstops and prepayments that Huang personally assembled can be run by committee — and nobody has tested that yet.

Should an interim CFO be allowed to sign new AI investments before a permanent CEO is named? Write to us via our contact page.

Get SiliconNoon in your inbox

One email a day with the AI stories that matter — what happened, the numbers, and why it matters. No spam, unsubscribe in one click.

Read more

Fractile raises $720M at $7B, six times its May valuation

Fractile raises $720M at $7B, six times its May valuation

Fractile, the Bristol- and London-based chip designer, has closed a $720 million Series C at a $7 billion post-money valuation, according to its investor Balderton. That is roughly six times the $1.1 billion it was worth in May, according to the Financial Times, when it raised a $220 million Series B — and it makes Fractile one of the most valuable technology companies in Britain, with a first commercial chip still a year from data centres. The round was co-led by Thrive Capital, Founders…

Deep Dive — OpenAI's $30B round fell to $20B at a $1.25T valuation

Deep Dive — OpenAI's $30B round fell to $20B at a $1.25T valuation

In September, OpenAI went to investors asking for $30 billion at a $1.4 trillion valuation. On Thursday night the Financial Times reported the target is now about $20 billion at $1.25 trillion, because some investors cut the size of their commitments amid the AI selloff, according to the FT. Hours earlier, the Wall Street Journal had put a number on how far the round has actually got: roughly $11 billion committed, less than half the original target, the Journal reported. That is the single…

OpenAI plans to buy Google TPUs as Oracle starts selling them

OpenAI plans to buy Google TPUs as Oracle starts selling them

For ten years the TPU was the one advantage Google refused to sell outside its own walls. This week it sold the distribution, and within a day the buyer on the other end was Nvidia's largest customer. Oracle will offer Google's tensor processing units through Oracle Cloud Infrastructure starting in calendar 2027, the first time TPUs will run on a hyperscale cloud other than Google's own, Oracle announced at its AI World conference.

Deep Dive — AI neoclouds hit a credit wall as SpaceX borrows $40B

Deep Dive — AI neoclouds hit a credit wall as SpaceX borrows $40B

On Sunday evening, Elon Musk announced on X that SpaceX had closed $40 billion of debt to buy Nvidia GPUs, and used the announcement to deliver a verdict on his competitors: the debt markets have "almost entirely shut off access to money for any company that doesn't have substantial revenue independent of AI data centers," and "the 200+ neoclouds are going to see guppy-level death and consolidation," he wrote. The $40 billion is the number everyone will quote.

Jensen Huang says no 2029 memory glut. The HBM numbers say 2027

Jensen Huang says no 2029 memory glut. The HBM numbers say 2027

Jensen Huang spent Thursday telling investors the memory cycle has years left in it. The strongest evidence that he is partly right is also the strongest evidence that the bill arrives well before 2029: chipmakers are already redesigning accelerators to use less high-bandwidth memory. What he actually said Asked about the memory selloff on the sidelines of Nvidia's GTC Berlin conference, Huang gave CNBC's Arjun Kharpal the sentence that will get quoted: "One of the things that's really…

Deep Dive — Alibaba wants to own the whole stack. Memory says no.

Deep Dive — Alibaba wants to own the whole stack. Memory says no.

The most important number Alibaba put on stage in Hangzhou this week was not ten trillion. Alibaba opened its annual Apsara Conference on Tuesday with a plan that is unusually easy to restate, because the company said it in one sentence: it will train Qwen models of five trillion to ten trillion parameters, run them on its own accelerators wired into clusters of up to 500,000 cards, and operate more than 20 gigawatts of data centre capacity worldwide by 2032.